Wikipedia as a reputation system

After speaking to Tantek Celik (thanks Tantek for the nice pocket cheat sheet on microformats at supernova!) and Dr John Breslin , I am increasingly becoming a fan of Micorformats .

Microformats offer significant potential for a number of reasons

a) We are seeiing increasing support from vendors for microformats – Open Social with FOAF, sixapart with technologies like OpenID, the Microformats hCard and XFN, and FOAF , the RDF autdiscovery firefox plugin among others ..

b) Microformats are key drivers behind data portability and social network portability

c) Microformats lend themselves to the mobile web because any format that fosters bite sized content and social networks is beneficial for the mobile web

In this blog, I am going to explore something beyond Microformats .. Let’s call it a ‘Wikipedia reputation system’ for the lack of a better word .. And I am seeking to address two different problems –

a) With Wikipedia – we have a reputation problem – i.e. an ‘expert’ who may know more may not be actually represented on the Wikipedia article. But – Is this a really a problem? I.e. how do you say who is the expert? Will the views of a few keep knowledge closed to the many?

b) With the semantic web – we have a different problem. Specifically – who will do the semantics for the semantic web and why?(in other words, add semantic information to the Web so that it can be program readable)

But let’s first start with a site called Naymz – which aims to be a reputation site on the web However, such initiatives have a problem. If my ‘reputation’ depends on updating someone else’s site – then basically it means – If I win the Nobel prize BUT forget to update that site, then my reputation is low – irrespective of my achievements

There are many variants of this theme but they have the same ‘Nobel prize’ problem – and that’s why most people who already have a good reputation will not join a site like this since the good reputation there depends on how many people you invite etc etc i.e. contributing to the site.

Essentially, it is Google bait .. (*Your free premium placement will typically appear on the first page of Google results in the ‘Sponsored Links’ section for searches on your name. Placements are limited to $10 in spend. We reserve the right to remove or discontinue placements at any time.) And you can even pay to repair your reputation

All this CAN be useful to people (for example visibility on Google)– so Naymz may indeed be a good business but in my view, that’s not a true reputation system

However, that does not mean we don’t need a global reputation system ..

One possibility for a global reputation system is Wikipedia with microformats

Wikipedia could be improved by a reputation system .. However, that reputation system has to be emergent ..

One solution( which could utilize Microformats) is to consider an iterative refinement of an article i.e. start with an article but keep refining it. Wikipedia does this anyway – but it does not lend itself to that mechanism. So, I am saying more of a threaded discussion from which expertise would emerge

This is almost like making the ‘discuss’ page of Wikipedia more visible and then putting people who made the best changes on to the main page i.e. experts will emege by their actions in the discussions when their changes have been accepted(a reputation system) from their activity on a globally neutral system(unlike ebay which is a closed, non portable reputation).

At the moment Wikipedia does not act as a reputation system i.e. the people who make the changes are not visible on Wikipedia and are not treated as experts as such

So, if Wikipedia becomes a defacto reputation system – then it becomes more interesting.

Now, add microformats to the mix – and we solve two problems at one stroke .. Wikipedia gets a richer framework since people can be motivated to add contextual information to make their content searchable (and that act directly benefits them anyway) and microformats get their semantics .. and the world gets an open reputation system

Why Wikipedia? Because we need something global, independent and neutral to which we all contribute content and contributing that content with semantics will help Wikipedia with its reputation problem and the semantic web with its semantic motivation problem. And we, as consumers could get a global reputation system

As usual thoughts welcome


  1. André Luís says:

    I could see this work if Wikipedia goes through the trouble of implementing a trustworthy algorithm for this. Simply presenting the top N contributors for any page woudn’t cut it, as it could probably be played. Maybe if they included a delay, say a week, would give time for any vandalism attack to be reverted by the ever-watching wikipedians. Then you’d be able to rely on the information of said group of hcards. :)
    Still, it’s a very valid idea. This could be done by including hCards with one rel-tag (relating to the subject of the page), pointing towards the profile of the user with the property UID. (this points to a page with a far more detailed hcard of the user) The usage of rel=”me” is arguable…
    But if you could summarize the whole set of rel-tags in the user profile page, you could even share the reputation by consolidating identity urls! With crossed rel=”me” links you could do this.
    Interesting concept… Microformats provide the vocabulary to get it across, but wikipedia needs to put in the trust part. ;)

  2. Ajit Jaokar says:

    Thanks Andre for your insightful comments. Often when I put out a blog like this – it is all very conceptual and experimental – so I am glad you like it and I like your feedback as well. kind rgds Ajit

  3. Sadly wikipedia is not interested in emergent authority. As Jimmy Wales told me if you want an entry in wikipedia “get better PR” – it counts links from offline journals…. as authorities, but not web people.